Subscribe to the gold package and get unlimited access to Shamra Academy
Register a new userDFOGraph: An I/O- and Communication-Efficient System for Distributed Fully-out-of-Core Graph Processing
369
0
0.0
(
0
)
Added by
Jiping Yu
Publication date
2021
fields
Informatics Engineering
and research's language is
English
Ask ChatGPT about the research
No Arabic abstract
With the magnitude of graph-structured data continually increasing, graph processing systems that can scale-out and scale-up are needed to handle extreme-scale datasets. While existing distributed out-of-core solutions have made it possible, they suffer from limited performance due to excessive I/O and communication costs. We present DFOGraph, a distributed fully-out-of-core graph processing system that applies and assembles multiple techniques to enable I/O- and communication-efficient processing. DFOGraph builds upon two-level column-oriented partition with adaptive compressed representations to allow fine-grained selective computation and communication, and it only issues necessary disk and network requests. Our evaluation shows DFOGraph achieves performance comparable to GridGraph and FlashGraph (>2.52x and 1.06x) on a single machine and outperforms Chaos and HybridGraph significantly (>12.94x and >10.82x) when scaling out.
rate research
Read More
Recent studies showed that single-machine graph processing systems can be as highly competitive as cluster-based approaches on large-scale problems. While several out-of-core graph processing systems and computation models have been proposed, the high disk I/O overhead could significantly reduce performance in many practical cases. In this paper, we propose GraphMP to tackle big graph analytics on a single machine. GraphMP achieves low disk I/O overhead with three techniques. First, we design a vertex-centric sliding window (VSW) computation model to avoid reading and writing vertices on disk. Second, we propose a selective scheduling method to skip loading and processing unnecessary edge shards on disk. Third, we use a compressed edge cache mechanism to fully utilize the available memory of a machine to reduce the amount of disk accesses for edges. Extensive evaluations have shown that GraphMP could outperform state-of-the-art systems such as GraphChi, X-Stream and GridGraph by 31.6x, 54.5x and 23.1x respectively, when running popular graph applications on a billion-vertex graph.
Analyzing massive complex networks yields promising insights about our everyday lives. Building scalable algorithms to do so is a challenging task that requires a careful analysis and an extensive evaluation. However, engineering such algorithms is often hindered by the scarcity of publicly~available~datasets. Network generators serve as a tool to alleviate this problem by providing synthetic instances with controllable parameters. However, many network generators fail to provide instances on a massive scale due to their sequential nature or resource constraints. Additionally, truly scalable network generators are few and often limited in their realism. In this work, we present novel generators for a variety of network models that are frequently used as benchmarks. By making use of pseudorandomization and divide-and-conquer schemes, our generators follow a communication-free paradigm. The resulting generators are thus embarrassingly parallel and have a near optimal scaling behavior. This allows us to generate instances of up to $2^{43}$ vertices and $2^{47}$ edges in less than 22 minutes on 32768 cores. Therefore, our generators allow new graph families to be used on an unprecedented scale.
Core decomposition is a fundamental graph problem with a large number of applications. Most existing approaches for core decomposition assume that the graph is kept in memory of a machine. Nevertheless, many real-world graphs are big and may not reside in memory. In the literature, there is only one work for I/O efficient core decomposition that avoids loading the whole graph in memory. However, this approach is not scalable to handle big graphs because it cannot bound the memory size and may load most parts of the graph in memory. In addition, this approach can hardly handle graph updates. In this paper, we study I/O efficient core decomposition following a semi-external model, which only allows node information to be loaded in memory. This model works well in many web-scale graphs. We propose a semi-external algorithm and two optimized algorithms for I/O efficient core decomposition using very simple structures and data access model. To handle dynamic graph updates, we show that our algorithm can be naturally extended to handle edge deletion. We also propose an I/O efficient core maintenance algorithm to handle edge insertion, and an improved algorithm to further reduce I/O and CPU cost by investigating some new graph properties. We conduct extensive experiments on 12 real large graphs. Our optimal algorithm significantly outperform the existing I/O efficient algorithm in terms of both processing time and memory consumption. In many memory-resident graphs, our algorithms for both core decomposition and maintenance can even outperform the in-memory algorithm due to the simple structures and data access model used. Our algorithms are very scalable to handle web-scale graphs. As an example, we are the first to handle a web graph with 978.5 million nodes and 42.6 billion edges using less than 4.2 GB memory.
The industry and academia have proposed many distributed graph processing systems. However, the existing systems are not friendly enough for users like data analysts and algorithm engineers. On the one hand, the programing models and interfaces differ a lot in the existing systems, leading to high learning costs and program migration costs. On the other hand, these graph processing systems are tightly bound to the underlying distributed computing platforms, requiring users to be familiar with distributed computing. To improve the usability of distributed graph processing, we propose a unified distributed graph programming framework UniGPS. Firstly, we propose a unified cross-platform graph programming model VCProg for UniGPS. VCProg hides details of distributed computing from users. It is compatible with the popular graph programming models Pregel, GAS, and Push-Pull. VCProg programs can be executed by compatible distributed graph processing systems without modification, reducing the learning overheads of users. Secondly, UniGPS supports Python as the programming language. We propose an interprocess-communication-based execution environment isolation mechanism to enable Java/C++-based systems to call user-defined methods written in Python. The experimental results show that UniGPS enables users to process big graphs beyond the memory capacity of a single machine without sacrificing usability. UniGPS shows near-linear data scalability and machine scalability.
Recent studies showed that single-machine graph processing systems can be as highly competitive as cluster-based approaches on large-scale problems. While several out-of-core graph processing systems and computation models have been proposed, the high disk I/O overhead could significantly reduce performance in many practical cases. In this paper, we propose GraphMP to tackle big graph analytics on a single machine. GraphMP achieves low disk I/O overhead with three techniques. First, we design a vertex-centric sliding window (VSW) computation model to avoid reading and writing vertices on disk. Second, we propose a selective scheduling method to skip loading and processing unnecessary edge shards on disk. Third, we use a compressed edge cache mechanism to fully utilize the available memory of a machine to reduce the amount of disk accesses for edges. Extensive evaluations have shown that GraphMP could outperform existing single-machine out-of-core systems such as GraphChi, X-Stream and GridGraph by up to 51, and can be as highly competitive as distributed graph engines like Pregel+, PowerGraph and Chaos.
Log in to be able to interact and post comments
comments
Fetching comments
Sign in to be able to follow your search criteria
